The first T. brucei protein identified to harbor methylated arginine residues was RBP16, a protein involved in mitochondrial RNA processing (Hayman and Read, 1999; Pelletier and Read, 2003). Three arginine residues, Arg-78, Arg-85 and Arg-93, are part of the RGG domain of RBP16 and methylation influences RBP16-RNA interactions (Pelletier et al., 2000; Miller and Read, 2003). Of these, only Arg-93 is constitutively methylated, while methylation of Arg-78 or Arg-85 appears to be mutually exclusive.

In vitro methylation assays using recombinant RBP16 and T. brucei procyclic whole cell protein extracts in the presence of classic substrates of type I and type II PRMTs indicated RBP16 is methylated by a trypanosome type I PRMT (Pelletier et al., 2001). The search for the type I PRMT that methylates RBP16 led to the identification of a protein 51% identical to the human PRMT1, thus named TbPRMT1, whose knockdown (KD) abolished RBP16^Arg78 and RBP16^Arg85 methylation (Pelletier et al., 2005; Goulah et al., 2006). Arg-93 remains methylated, likely due to the action of another PRMT. Curiously, TbPRMT1 is mostly present in the cytoplasm, suggesting RBP16 might be methylated before import into the mitochondrion (Fisk et al., 2010).

In T. brucei, RBP16 forms complexes of various sizes, but most notably 5S and 11S complexes; the latter likely represents the proteinaceous 5S complex bound to RNAs. Curiously, cells depleted for TbPRMT1 or expressing the R78K, R85K and R93K triple mutant RBP16 formed a 5S complex composed of RBP16 bound only to mitochondrial guide RNAs (gRNAs), implicit in RNA editing, but not mRNAs (Figure 1B). Accordingly, non-methylatable RBP16 has an increased affinity for gRNAs, yet displays lower affinity for mitochondrial mRNAs (Goulah and Read, 2007). It is, however, unknown which proteins interact with RBP16 and whether mitochondrial mRNAs bind directly to RBP16 in the complex. Nonetheless, expression of non-methylatable RBP16 has been associated with destabilization of NADH dehydrogenase subunit 4 mRNA, whose quantity is also lower in TbPRMT1-KD cells, though TbPRMT1-knockdown also impacts levels of other mRNAs (Goulah et al., 2006). The effects of arginine methylation by TbPRMT1 are not limited to the mitochondrion. DRBD18 is a cytoplasmic RBP whose methylation state leads to different protein complex formation and alters mRNA expression. RNAs stabilized by methylated DRBD18 are less stable in the absence of TbPRMT1 or when non-methylatable DRBD18 is expressed; the opposite is true for RNAs destabilized by DRBD18 (Lott et al., 2015). In fact, TbPRMT1 knockout has a broad, complex effect on mRNP associations, which impacts cell metabolism, particularly energy production pathways, as well as stress granule formation, and results in reduced in vitro T. brucei growth and virulence in mice (Kafková et al., 2018).

Like TbPRMT1, TbPRMT3 was also identified as a potential type I PRMT and knockdown of either TbPRMT1 or TbPRMT3 reduced aDMA levels in the cells. Interestingly, the reduction in the protein level of either was accompanied by a reduction of the other, suggesting an interdependent stability between TbPRMT1 and TbPRMT3 (Pelletier et al., 2005; Lott et al., 2014).

However, TbPRMT3 is inactive in vitro (Kafková et al., 2017), and its primary sequence lacks conserved residues in THW and double E loops, which are typically well conserved and responsible for substrate binding and positioning, respectively (Tewary et al., 2019). Structural data showed that although TbPRMT3 retains the four canonical PRMT domains (N-terminus, SAM-binding Rossman fold domain, dimerization arm and β-barrel domain; Figure 2A), it lacks a crucial 3₁₀ α-helix in the Rossman fold, which alters the dimerization interface and precludes SAM binding (Kafková et al., 2017; Hashimoto et al., 2020). Importantly, T. cruzi PRMT3 also lacks conserved THW and double E loops.

Functional studies have shown that TbPRMT3 is essential for TbPRMT1 stability and activity, establishing it as a “prozyme” or “pseudoenzyme” that supports the catalytically active TbPRMT1. TbPRMT1 and TbPRMT3 have, thus, been renamed as TbPRMT1^ENZ and TbPRMT1^PRO, respectively (Kafková et al., 2017; Hashimoto et al., 2020) (Figure 1A). Interestingly, when amino acids 41–52 were removed from TbPRMT1^PRO, methyltransferase activity was lost as the complex could no longer bind substrates. Although enzymatically inactive, TbPRMT1^PRO contributes to substrate recognition (Hashimoto et al., 2020).

Unlike mammalian counterparts, TbPRMT1^ENZ and TbPRMT1^PRO interact via a hydrophobic interface to form a ring-like heterodimeric structure (Figure 2B), similar to homodimeric PRMTs. Two TbPRMT1^ENZ‒TbPRMT1^PRO heterodimers together form the functional heterotetramer (Hashimoto et al., 2020). Curiously, TbPRMT1^PRO can homodimerize in solution, albeit minimally (Kafková et al., 2017), whereas TbPRMT1^ENZ is challenged to form homodimers (Hashimoto et al., 2020), supporting structural dependence on TbPRMT1^PRO.

It is possible that TbPRMT1^PRO (alone or as a homodimer) performs moonlighting functions in the cell, as TbPRMT1^PRO transcripts were slightly increased upon DNA damage induction, while TbPRMT1^ENZ transcripts were reduced. TbPRMT1^PRO has also been shown to bind mRNA both in vitro and in vivo independent of tetramer formation (Lueong et al., 2016; Stortz et al., 2017; Kafková et al., 2018). The biological relevance of proposed moonlighting properties remains unclear.

TbPRMT5 is the only type II PRMT expressed by the parasite, yet the least studied T. brucei PRMT and the only one whose structure remains unsolved. TbPRMT5 is by far the largest T. brucei PRMT due to a very long N-terminal region, the biological relevance of which is unknown (Figure 2A). In vitro, recombinant TbPRMT5 displays a broad substrate specificity that includes RBP16 (Pasternack et al., 2007). Whether TbPRMT5 methylates RBP16 in vivo is unknown, but if so, TbPRMT5 may be responsible for constitutive methylation of RBP16^Arg93. Moreover, the fact that recombinant TbPRMT5 is active suggests that, unlike mammalian PRMT5, it does not require a co-factor to function. Further to this, no homologues of mammalian PRMT5 methylosome components were identified in TbPRMT5 immunoprecipitation experiments (Pasternack et al., 2007).

Under native conditions, TbPRMT5 is found in different protein complexes, with sizes ranging between 150 and 700 kDa (Pasternack et al., 2007). Strikingly, TbPRMT5 interacts with Kinetoplastid-specific proteins, suggesting importance in Kinetoplastid-specific pathways (Pasternack et al., 2007). Further studies are necessary to verify TbPRMT5 substrates and determine consequences of PRMT5-dependent methylation. In addition, sequence data indicates that the currently uncharacterized Leishmania PRMT5 is much larger (>1,000 amino acids) than its orthologues in humans and T. brucei (637 and 784 amino acids, respectively) due to a much longer N-terminus that contains no known conserved functional domains. Elucidating the 3D structure of at least one trypanosomatid PRMT5 will lend insight into potential functional roles of this N-terminus.

TbPRMT6 is a type I PRMT that displays a narrow substrate specificity. Accordingly, TbPRMT6 knockdown does not visibly alter the cellular arginine methylation profile (Lott et al., 2014). Results suggest that protein targets of TbPRMT6 methylation are important for T. brucei cellular replication, given TbPRMT6-KD caused a mild growth defect in vitro and led to the appearance of aberrant cells (Fisk et al., 2010).

TbPRMT6 is expressed by both T. brucei procyclic and bloodstream forms at equal levels and, despite being primarily cytoplasmic, it interacts with several histones. TbPRMT6 also co-purifies with proteins involved in nucleocytoplasmic transport and RNA processing, indicating it might be important in controlling nucleic acid metabolism and transport (Fisk et al., 2010). Disruption of these processes is known to cause growth defects in vitro and TbPRMT6 transcription is reduced in cells exposed to DNA damage (Stortz et al., 2017).

TbPRMT6 contains the four canonical PRMT domains (Figures 2A,C), and holds unique sequence and structural features that appear conserved across the Kinetoplastids. Secondary structure analysis of type I PRMTs indicates that TbPRMT6 contains four insert regions and a truncated C-terminus. The insertions seem to extend or introduce additional α-helices (Wang et al., 2014b), indicating potential relevance to methyltransferase activity and/or regulation, as well as substrate selection. The exact role of these peculiarities is still to be determined.

Importantly, the 3D structure of apo-TbPRMT6 complexed with S-adenosylhomocysteine (SAH) revealed that substrate binding remodels the active site to allow correct positioning of the target arginine residue. These conformational changes involve residues that are conserved in type I PRMTs, including His318 in the THW loop and Glu142 in the Double E loop, suggesting this feature may be conserved among type I enzymes (Wang et al., 2014b).

Kinetoplastids are the only unicellular eukaryotes known to express a PRMT7 homolog. Whereas the mammalian PRMT7 polypeptide contains two copies of the core PRMT fold, which interact with each other to form an intramolecular- or pseudo-dimer (Cura et al., 2014), TbPRMT7 contains only a single active site and is almost half the size of human PRMT7, although much more active (Fisk et al., 2009). TbPRMT7 is a cytoplasmic enzyme expressed by both long-slender bloodstream and procyclic form Trypanosoma brucei, and at least in the latter, it forms different macromolecule complexes (Fisk et al., 2009, Fisk et al., 2010). The composition of these complexes is still unknown, but it likely contains RNA, as TbPRMT7 can bind RNA in vitro and in vivo (Lueong et al., 2016; Kafková et al., 2018).

Curiously, knockdown of TbPRMT7 only minimally affects the MMA profile, which might be due to an increase in the monomethylation activity of other PRMTs (Lott et al., 2014). Simultaneous knockdown of TbPRMT7 and TbPRMT1 reduced both MMA and aDMA levels in the cells, whereas the knockdown of TbPRMT1 alone caused an accumulation of MMA. All evidence suggests that TbPRMT7 generates monomethylated substrates for other PRMTs and that TbPRMT1 activity can compensate for reduction of TbPRMT7 activity (Lott et al., 2014). Accordingly, recombinant TbPRMT7 displays broad substrate specificity in vitro, which includes proteins known to be methylated by TbPRMT1 and TbPRMT5, such as RBP16 and TbRGG1 (Fisk et al., 2009). It is therefore possible that TbPRMT7 is also involved in RBP16^Arg93 constitutive methylation.

The TbPRMT7 3D structure showed the expected four canonical domains of PRMTs (Figures 2A,D). Similar to the other Kinetoplastid PRMTs, homodimerization is facilitated by hydrophobic interactions between a dimerization arm and the SAM-binding domain of the other subunit (Wang et al., 2014a). Extensive mutations on the dimerization arm abolished dimerization, leaving only residual methyltransferase activity (Debler et al., 2016).

Each TbPRMT7 monomer can bind SAM and arginine substrate molecules (Wang et al., 2014a; Debler et al., 2016), though the arginine substrate binding pocket appears to be significantly narrower than those of type I and II PRMTs, consistent with its ‘monomethylation only’ profile. In fact, conserved residues present in the double E and THW loops restrict TbPRMT7 to monomethylation. An E181D mutant was able to catalyze aDMA (Debler et al., 2016), while an E181D/Q329A double mutant generated sDMA (Jain et al., 2016). Simulation studies also indicate that E172 and Q329 are crucial for proper substrate orientation and facilitating the reaction mechanism (Thakur et al., 2019). Furthermore, a F71I mutant was able to form dimethylated products (Jain et al., 2016; Cáceres et al., 2018). These data showed the importance of the double E loop with each of E172 and E181 forming two hydrogen bonds to the guanidino group of the substrate arginine (Wang et al., 2014a) (Figure 2E).